本文对分布式迭代随机大系统给出了定性分析．对时不变线性、时变线性和非线性迭代随机大系统建立了关联均方收敛性判据．在所有孤立子系统都收敛的情形，只需选择子系统中适当的乘子，迭代随机大系统就能对任意互连项和在结构扰动下保持均方收敛性． This paper gives a qualitative analysis of distributed iterative stochastic large-scale system. The mean-square convergence criterion for time invariant linear, time-varying linear and nonlinear iterative stochastic systems is established. When all the isolated subsystems converge, simply select the appropriate multipliers in the subsystem and iterate the stochastic large-scale system to preserve the mean-square convergence of any interconnected terms and structural perturbations.